My last post, posed the question: are predictive models still relevant in today’s economic climate? As expected it generated a fair amount of discussion both on and off the blog. Thank you for all of your comments. Let’s continue the dialogue.

A Quick Poll…

Scoring isn’t dead, but the type matters

In this market, it’s fair to be skeptical, but abandoning all forms of scoring completely would likely do more harm than good. Having some method of consistently evaluating and rank ordering customers remains the best, most efficient way to manage these relationships for both upfront credit decisioning and ongoing portfolio monitoring and collections activities. However, carefully choosing the type of model and conducting some near-term validation is certainly warranted. Based on discussions with a variety of credit and collections professionals, this market appears to be leaning toward a more point-in-time view of their prospects and current customers such as:

• financial statement-based scoring (for both public and private companies). Many reported more sharing of private financial statements than prior to the economic crisis.
• more judgmental, short-term payment history scores that focus on the current state of the business rather than trying to predict future payment behavior using models that may no longer be valid. Some of the comments from the last post landed on both sides of the debate on whether predictive models are experiencing “model drift” and under or over predicting risk or are these models holding up and still predicting accurately? (we welcome your continued comments on this debate)
• finding methods to compare internal payment behavior (how they pay me) with external data (how they pay others)

Get a complete view of your portfolio

We also heard from credit pros that they are increasing turning to a variety of information sources to get a better picture of how they are getting paid vs. their peers and then reporting this information back to senior management. This “better/worse” than the benchmark analysis can prove quite enlightening for those outside of the day-to-day credit and collections function. This can take many forms:

• comparing your own payment experience against an industry benchmark
• comparing your own payment experience against your credit group benchmark
• comparing your own payment experience against the broader market

Resources to help you guide your use of scoring into today’s economic climate

Trying to rate and analyze large volumes of both prospective and existing customers, scoring is still a must, especially as many companies have cut staff. Below are some helpful resources* that you can use to guide your assessment of the various options on the market today:

• Credit Today: Their Credit Scoring Central is probably one of the largest libraries on this topic in a user-friendly, “plain English” format that often includes case studies from credit professionals in many industries.
• Credit Research Foundation: An early leader and educator on the topic that offers a wealth of research papers and industry studies. The CRF also publishes a variety of publications as well as hosts industry events, seminars and best practice workshops that often cover various forms of scoring.
• NACM: Business Credit Magazine offers a Resource Library with past coverage of commercial credit scoring articles.
• Financial Insights: Dana Wiklund and an analyst with advisory firm, IDC recently published this video presentation: “Why Now Is The Time To Validate Your Models“

*Please note that some of these are available for a fee or subscription service.

Let’s keep the dialogue going

What are your thoughts on how scoring should be applied and what types of scoring you are using today?

Posted in Information, Predictive Models, Scoring
Permalink: http://blog.cortera.com/2009/04/27/what-kind-of-scoring-is-best-for-you/